Groundwater pumping from Bicycle Groundwater Basin (referred to as Bicycle Basin) in the Fort Irwin National Training Center, California, began in 1967. From 1967 to December 2010, about 46,000 acre-feet of water had been pumped from the basin and transported to the Irwin Basin. During this time, not only did water levels in the basin decline by as much as 100 feet, the quality of the groundwater pumped from the basin also deteriorated in some wells.

The U.S. Geological Survey collected geohydrologic data from existing wells, test holes, and 16 additional monitoring wells installed at 6 sites in Bicycle Basin during 1992–2011 to determine the quantity and quality of groundwater available in the basin. Geophysical surveys, including electrical, gravity, and seismic refraction surveys, were completed to help determine the geometry of the structural basin, delineate depths to the interface between Quaternary and Tertiary rocks, map the depth to the water table, and used to develop a geohydrologic framework and groundwater-flow model for Bicycle Basin.

Water samples were used to determine the groundwater quality in the basin and to delineate potential sources of poor-quality groundwater. Analysis of stable isotopes of oxygen and hydrogen in groundwater indicated that present-day precipitation is not a major source of recharge to the basin. Tritium and carbon-14 data indicated that most of the groundwater in the basin was recharged prior to 1952 and had an apparent age of 15,625–39,350 years. Natural recharge to the basin was not sufficient to replenish the groundwater pumped from the basin. Interferograms from synthetic aperture radar data (InSAR), analyzed to evaluate land-surface subsidence between 1993 and 2010, showed 0.23 to 1.1 feet of subsidence during this period near one production well north of Bicycle Lake (dry) playa.

A groundwater-flow model of Bicycle Basin was developed and calibrated using groundwater levels for 1964–2010, and a subsidence model using land-surface deformation data for 1993–2010. Between January 1967 and December 2010, the simulated total recharge from precipitation runoff and underflow from adjacent basins was about 5,100 acre-feet and pumpage from the Bicycle Basin was about 47,000 acre-feet of water. Total outflows exceeded natural recharge during this period, resulting in a net loss of about 42,100 acre-feet of groundwater storage in the basin.

The Fort Irwin National Training Center is considering various groundwater-management options in the Bicycle Basin. The groundwater-flow model was used to (1) evaluate changes in groundwater levels and subsidence with the addition of capture and recharge of simulated runoff in retention basins (scenario 1) for predevelopment through 2010; (2) simulate a base case (scenario 2) for reference; and (3) compare projections of alternative future pumping strategies for 2011–60 (scenarios 3–5).

Model results from the runoff-capture simulation (scenario 1) indicated that total recharge, including runoff captured using retention basins, locally increased water levels, which partially offset, but did not mitigate, groundwater depletion associated with pumping. Groundwater-storage depletion in scenario 1 was about 14 percent less than without runoff capture. Simulated-drawdown results in model layer 1 in the eastern part of the basin indicated that, because of the captured runoff, simulated heads were as much as 100 feet higher in December 2010 than prior to the onset of development in 1967. In contrast, simulated drawdown for model without runoff capture indicated that, without captured runoff, simulated heads for December 2010 in this area were 80–90 feet lower than during the predevelopment period. Subsidence was mitigated slightly in scenario 1 compared to without runoff capture; the largest decrease in subsidence at observation sites was about 0.07 feet.

Scenario 2 results indicated that simulated drawdown in model layer 6 at the end of 2060 ranged from about 46 to 135 feet. Subsidence at observation sites at the end of 2060 ranged from 0.83 to 2.8 feet. Reducing the base case pumping rate by 25 percent in the existing production well in the subsidence area and redistributing the pumpage to the other two production wells (scenario 3) resulted in a reduction of drawdown in the subsidence area compared with the base case (scenario 2). The difference in subsidence at the end of 2060 between scenario 3 and the base case was small (less than 0.07 feet) for all observation locations. Repeating the simulation scenario 3 but additionally reducing the basin-wide pumpage by 3 percent per year from 2011 through 2020 (scenario 4) resulted in about 60 feet less drawdown in the subsidence area than for the base case. Subsidence at observation sites ranged from 0.12 to 0.43 feet less than for the base case. Reducing the pumpage in the existing production well in the subsidence area to zero, while continuing the base case pumping rate in the other two existing production wells (scenario 5), resulted in more than 100 feet less drawdown in the subsidence area than in the base case. The simulated subsidence at the end of 2060 ranged from about 0.19 to about 1.16 feet less than in the base case, indicating that the discontinuation of pumpage at well 14N/3E-14P1 would result in substantially reduced subsidence.

Overall, continued water-table declines dewatered the more productive upper layers of the aquifer, causing more groundwater to be withdrawn from deeper, lower yielding layers and resulting in faster declines in the water table and greater vertical gradients in the future. If the water table declines into the perforated intervals of wells, increased maintenance costs and altered well-water quality could potentially result. Water-management scenario 1 indicated that adding managed recharge resulted in a modest decrease (4–5 percent) in the rate of subsidence compared with historical conditions. The reduction of pumpage in the area of subsidence and redistribution of the amount of reduced pumpage to wells in other parts of the basin (scenario 3) resulted in modest decreases (5–6 percent) in the rate of subsidence compared with continuation of historical pumpage. Including either a basin-wide reduction of pumpage of 3 percent annually (cumulatively, a 24 percent decrease) as well as redistribution of the pumpage (scenario 4) or discontinuing pumping in the area of greatest subsidence (scenario 5), however, demonstrated that subsidence could be reduced substantially (22–26 percent for scenario 4 and 62–68 percent for scenario 5) compared with subsidence for continuing historical pumpage.